Causality is the relationship between an event (the cause) and a second event (the effect), where the second event is understood as a consequence of the first.

A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty. Uncertainty is the essence of a startup. A startup operates in an uncertain environment: both the problem to be solved and the product to build are unknown. This is why learning is the vital function of a startup. 

Learning means updating our theories about reality, and these theories are expressed as cause-effect relationships. The problem is that our understanding of causality is often broken, and since our decisions are all influenced by our assumptions on causality, they have the potential to sway them.

In this post I investigate the problems of cause-effect relationships and show how we can overcome this obstacle.

The problem with causality

Learning is all about understanding why things happen and why some events lead to other events. Learning is about understanding the casual link between things. 

But this casual link doesn’t exist in reality; it isn’t a property of things. We cannot perceive cause and effect, but as the Scottish philosopher David Hume once said, we develop a habit of mind where we come to associate two types of object and event, always contiguous and occurring one after the other. Contiguity and succession are what we call causality.

Now, we infer causality from sensory information, prior experience and innate knowledge. And here come our problems:

Limited information - Only limited, often unreliable, information is available; in complex situation, many interdependent factors affect outcomes. Sometimes we ignore how factors are linked between themselves, sometimes we ignore what factors affect a situation: we usually deal with partial information;

Limited capacity - Human mind has only limited capacity to evaluate and process the information that is available. Usually, our prior experience and knowledge teach us what to look at;

Attribution errors - Human beings have a tendency to over-value dispositional explanations for the observed behaviors of others while under-valuing situational explanations for those behaviors. For example, we are likely to attribute our success to our insights and skills while downplaying others’ success to good luck or other external factors.

How to overcome these obstacles?

David Hume suggested what he calls mitigated skepticism: since we all have limited experience, our conclusions should always be tentative, modest, reserved, cautious. Nonetheless, anyone “sensible of the strange infirmities of human understanding” should also strive for overcoming these obstacles, and here is my advice. 

Data-informed - Make data-informed, not data-driven decisions. Data give you half the picture: they measure the past, and cannot predict the future. Treat data with caution: you might have measured the wrong thing, you might have measured the right thing in the wrong way. 

Test your assumptions - When too much uncertainty prevents you from a full understanding of your problem, you have only one choice: test your hypotheses. In science, this is known as the scientific method. If you think you’re doing right, test your assumptions. Do it.

Stay flexible - When there is too much uncertainty, flexibility is the key to success. Don’t make long term commitments that harm your capability to adapt to a fast changing environment. Napoleon would order his troops to stay at most 1-day march distance so that they were in mutually supporting positions and able to come to the aid of each other in the event of concentration for battle or to ward of superior forces. Since the battle plan was formulated only when enemy’s intentions were clear, this formation gave the Frenches the flexibility to gain a strategic advantage over their enemies.

I don’t want to encourage risk-taking behavior. In 218 BC, Hannibal invaded Italy by land across the Alps. The task was daunting to say the least. Nobody thought it would be possible. But Hannibal knew it was possible, and did it. Intuition is a highly developed pattern recognition process. It have to be trained, and nothing can train intuition like experience. 

————————————————

Have you ever followed your gut instinct when odds were against you? Let me know your experience with a comment.